Scholarly record

NEURAL NETWORKS AS OPTIMIZATION TOOLS FOR FUEL CONSUMPTION

Costel Humelnicu

Abstract

The paper presents a neural network based methodology for prediction and optimization of internal combustion engine vehicles' fuel consumption, as a way to reduce air pollution. As it is well demonstrated lately, the engine fuel burning is one of the main factors that generate air pollution. Its impact on environment is rapidly increasing, following the increase of vehicles numbers. The best solution to reduce the air pollution is to use electric motor driven vehicles but these are still very expensive, with a low commercial rate. So, the optimization of current engines, by tuning the main parameters like power, torque, cylinder number etc. stands for an affordable solution. Due to many influencing parameters, it is difficult to predict the fuel consumption value based only on theoretical calculus. Neural network models allow the integration of experimental acquired data, taking this way into account the mutual influences between internal combustion engine parameters, leading to a more precise estimation of fuel consumption. Taking into account that the neural network architecture is directly linked to the modelled phenomenon, several networks were tested, in order to find the one suitable for this work?s goal. Once the best model is identified, predictions and optimization procedures can be performed.

Publication Impact Profile

• Citations
• Scopus - Citation Indexes: 3

• Captures
• Mendeley - Readers: 4

PLUMX see details

DOI resolver: doi.org/10.5593/sgem2018/4.2/s19.069

Publication details

Back to publication list